CardiacMeshDataset

class CardiacMeshDataset(mesh, mesh_numbers, timing_ms)

Cardiac mesh model assuming the continuous node-ordering:

[apex, n_points_per_ring(1st slice, 1st shell), n_points_per_ring(2st slice, 1st shell), …, n_points_per_ring(1st slice, 2nd shell)]

shell –> radial slice –> longitudinal

Parameters:

• mesh (Union[UnstructuredGrid, DataSet]) – pyvista mesh containing the field-array ‘displacement_{time}ms’ for each time specified in timing_ms

• mesh_numbers ((int, int, int)) – (shells, circ, long)

• timing_ms (ndarray) – array containing the times corresponding to the stored displacements

Methods:

compute_cylinder_coords(time_stamp)	Computes cylinder coordinates (transmural_depth, polar_angle, z-pos) for all mesh nodes.
compute_local_basis(time_stamp[, keys])	Computes the local coordinate system (radial, circumferential, longitudinal) for a left ventricle mesh.
evaluate_cellsize(points, connectivity)	Evaluates cell centers and cell volumes for the tetra-mesh given as points and connectivity.
evaluate_connectivity(n_theta, n_radial, ...)	Generates connectivity for tetrahedral UnstructuredGrid in vtk format Inputs: - n_theta : number of points along circumferential contour (n points for each ring) - n_radial : number of points along longitudinal direction - n_layers : number of points along radial direction
evaluate_cylinder_coordinates(...)	type positional_vectors: ndarray
from_list_of_arrays(initial_mesh, ...[, ...])	Loads a mesh and its displacements from a list of numpy files each containing the positional vectors for all points in initial_mesh.
from_list_of_meshes(list_of_files, timing, ...)	Loads vti/vtk files that contain the motion states of the same mesh, calculates the displacements from positional differences and returns a CardicMeshDataset instance containing the displacements
from_single_vtk(file_name, mesh_numbers[, ...])	Loads a vtk-file that contains the displacements
interpolate_mesh(mesh_points[, ...])	Function to linear interpolate points of LV meshes in circular, longitudinal and
refine(longitudinal_factor, ...)	Refines the cardiac mesh model assuming the continuous node-ordering:

Attributes:

mesh_numbers

(n_shells, n_points_per_ring, n_slices)

compute_cylinder_coords(time_stamp)

Computes cylinder coordinates (transmural_depth, polar_angle, z-pos) for all mesh nodes.

Parameters:

time_stamp (Quantity) – time Quantity used to determine reference mesh configuration

compute_local_basis(time_stamp, keys=('e_t', 'e_c', 'e_l'))

Computes the local coordinate system (radial, circumferential, longitudinal) for a left ventricle mesh. The local basis vectors per point are stored with specified keys

Parameters:

• time_stamp (Quantity) – time Quantity used to determine reference mesh configuration

• keys (Tuple[str, str, str]) – array names used to store (radial, circumferential, longitudinal) basis vectors

static evaluate_cellsize(points, connectivity)

Evaluates cell centers and cell volumes for the tetra-mesh given as points and connectivity.

Parameters:

• points (ndarray) – (#points, 3)

• connectivity (ndarray) – (#cells, 4)

Return type:

(ndarray, ndarray, ndarray)

Returns:

cell_center (#cells, 3), cell_volumes (#cells), point_weights (#points)

static evaluate_connectivity(n_theta, n_radial, n_layers)

Generates connectivity for tetrahedral UnstructuredGrid in vtk format Inputs:

• n_theta : number of points along circumferential contour (n points for each ring)

• n_radial : number of points along longitudinal direction

• n_layers : number of points along radial direction

Outputs:

• connectivity : connectivity array where row i contains the vertex id of points

defining tetrahedron with index i

Dr. Buoso Stefano (2020) email: buoso@biomed.ee.ethz

Parameters:

• n_theta (int) –

• n_radial (int) –

• n_layers (int) –

static evaluate_cylinder_coordinates(positional_vectors, n_shells, points_per_shell)

Parameters:

• positional_vectors (ndarray) – (-1, 3)

• n_shells (int) – (int)

• points_per_shell (int) – (int)

Returns:

np.ndarray (-1, 3) [radial, angle, relative z]

classmethod from_list_of_arrays(initial_mesh, mesh_numbers, list_of_files, timing, time_precision_ms=1)

Loads a mesh and its displacements from a list of numpy files each containing the positional vectors for all points in initial_mesh.

Raises:

FileNotFoundError - if not all files in given list of files exist

Parameters:

• initial_mesh (DataSet) – Initial mesh configuration

• mesh_numbers ((int, int, int)) – (shells, circ, long)

• list_of_files (List) – List of .npy containing the position of all mesh nodes per time

• timing (Quantity) –

• time_precision_ms (int) – number of decimals used in the field name and timepoints (e.g. displacements_0.01ms)

classmethod from_list_of_meshes(list_of_files, timing, mesh_numbers, field_key='displacement', time_precision_ms=1)

Loads vti/vtk files that contain the motion states of the same mesh, calculates the displacements from positional differences and returns a CardicMeshDataset instance containing the displacements

Raises:

FileNotFoundError - if not all files in given list of files exist

Parameters:

• list_of_files (List) – List of .vtk or .vti files each containing a mesh in which the displacement vectors for each node is stored under the name of specified by the argument field_key

• timing (Quantity) –

• mesh_numbers ((int, int, int)) – (shells, circ, long)

• field_key (str) – key to extract

• time_precision_ms (int) – number of decimals used in the field name and timepoints (e.g. displacements_0.01ms)

Return type:

CardiacMeshDataset

classmethod from_single_vtk(file_name, mesh_numbers, time_precision_ms=3)

Loads a vtk-file that contains the displacements

Parameters:

• file_name (str) –

• mesh_numbers ((int, int, int)) – (shells, circ, long)

• time_precision_ms (int) – number of decimals used in the field name and timepoints (e.g. displacements_0.01ms)

Returns:

static interpolate_mesh(mesh_points, points_per_ring=40, n_shells=4, n_slices=30, interp_factor_c=1, interp_factor_l=1, interp_factor_r=1)

Function to linear interpolate points of LV meshes in circular, longitudinal and

radial direction. Indexing of the points is assumed to adhere to the following structure:

0th element -> Apex 1:points_per_ring+1 -> inner most shell most apical slice. Repeated for n_slices. Repeated for n_shells.

Parameters:

• mesh_points (ndarray) – np.ndarray - (-1, 3) Flattened array of 3D positional vectors of mesh points.

• points_per_ring (int) – (int) number of points per ring (circumferential direction)

• n_shells (int) – (int) number of shells (radial direction)

• n_slices (int) – (int) number of slices (longitudinal direction)

• interp_factor_c (int) – (int) interpolation multiplier for circumferential direction

• interp_factor_l (int) – (int) interpolation multiplier for longitudinal direction

• interp_factor_r (int) – (int) interpolation multiplier for radial direction

Return type:

(ndarray, Tuple[int, int, int])

Returns:

np.ndarray - (n_shell_new, 1 + (n_theta_new*n_slices_new), 3), (n_theta_new, n_slices_new, n_shells_new)

refine(longitudinal_factor, circumferential_factor, radial_factor)

Refines the cardiac mesh model assuming the continuous node-ordering:

[apex, n_points_per_ring(1st slice, 1st shell), n_points_per_ring(2st slice, 1st shell), …, n_points_per_ring(1st slice, 2nd shell)]

shell –> radial slice –> longitudinal

Parameters:

• longitudinal_factor (int) – nl_new = f * nl_old

• circumferential_factor (int) – nc_new = n

• radial_factor (int) –

Return type:

CardiacMeshDataset

Returns:

New CardiacMeshDataset instance with refined mesh and new in

mesh: Union[UnstructuredGrid, DataSet]

Mesh containing the displacements for each time-point

mesh_numbers: Tuple[int, int, int]

(n_shells, n_points_per_ring, n_slices)

timing: ndarray

Time points corresponding to the mesh states given by the list of files in ms